Plot a multichannel signal to identify the frequency value

I have a multi-channel EEG data of size 64x1000, where 64 is the number of channels and 1500 time points. I already know that this signal is of 8Hz. I want to plot this signal to show that there is a peak 8Hz indicating the signal frequency.

How can I do it??

Answer 1

[Jesus Sanchez]

I would do a fast fourier transform for each channel (fft in matlab), so you have the frequential componentes of each channel. Then just use a function such as mesh, surf or pcolor. As these functions represent 3D surfaces, I would define X as channel, Y as frequency and Z as the value given by the FFT.

So in general lines the code will be something like this:

 nFFT = 2048; % I asked for 2048 frequency points because why not.
nChannels = 64;
data_freq =fft( multi-channel-EEG-data.',nFFT); 
% FFT is applied to columns, that is why it is neccesary to transpose them. 
x = 1:nChannels; 
y = 1:nFFT;
[X,Y]=meshgrid(x,y);
figure
mesh(X,Y,data_freq);
shading interp;

I did not try the code, so there are probably several mistakes, but if you follow this lines, it should work :). You might need to apply another .' to data_freq to represent it. Good luck!

[Riddhi Gala]

This doesn't work. I don't have a peak at 8Hz.

[Jesus Sanchez]

Can you upload your data? If there is a peak appearing somewhere different from 8 Hz, it could be that the plot is shifted. Moreover, maybe 2048 points for the FFT is too little. Try to do it directly with

 fft( multi-channel-EEG-data.');

and see if that works.